Design and Performance of the CDC Real-Time Reverse Transcriptase PCR Swine Flu Panel for Detection of 2009 A (H1N1) Pandemic Influenza Virus

ABSTRACT Swine influenza viruses (SIV) have been shown to sporadically infect humans and are infrequently identified by the Influenza Division of the Centers for Disease Control and Prevention (CDC) after being received as unsubtypeable influenza A virus samples. Real-time reverse transcriptase PCR (rRT-PCR) procedures for detection and characterization of North American lineage (N. Am) SIV were developed and implemented at CDC for rapid identification of specimens from cases of suspected infections with SIV. These procedures were utilized in April 2009 for detection of human cases of 2009 A (H1N1) pandemic (pdm) influenza virus infection. Based on genetic sequence data derived from the first two viruses investigated, the previously developed rRT-PCR procedures were optimized to create the CDC rRT-PCR Swine Flu Panel for detection of the 2009 A (H1N1) pdm influenza virus. The analytical sensitivity of the CDC rRT-PCR Swine Flu Panel was shown to be 5 copies of RNA per reaction and 10−1.3∼−0.7 50% infectious doses (ID50) per reaction for cultured viruses. Cross-reactivity was not observed when testing human clinical specimens or cultured viruses that were positive for human seasonal A (H1N1, H3N2) and B influenza viruses. The CDC rRT-PCR Swine Flu Panel was distributed to public health laboratories in the United States and internationally from April 2009 until June 2010. The CDC rRT-PCR Swine Flu Panel served as an effective tool for timely and specific detection of 2009 A (H1N1) pdm influenza viruses and facilitated subsequent public health response implementation.

Swine influenza is a common contagious respiratory disease of pigs caused by influenza A viruses (6,11). The first swine influenza virus (SIV), also referred to as classical SIV H1N1 virus, was isolated in the United States in 1930 and circulated in pigs for at least 60 years (22,24). From 1997 to 1998, triple-reassortant SIV (tr-SIV) H1N1, H1N2, H3N2, and H2N3 subtypes containing gene segments from human, avian, and swine influenza viruses were found to be circulating in North American (N. Am) swine populations (13,14,29,33). Genetic analysis showed that, in particular, the nucleoprotein (NP) gene of tr-SIVs was inherited from classical SIV, that the origins of the hemagglutinin (HA) and neuraminidase (NA) genes differed, and that those genes were most closely related to genes of classical SIV (H1N1), and recently isolated human (H1N2 and H3N2) or avian (H2N3) viruses (17,30,33) (Fig.  1). These findings provide compelling evidence that tr-SIVs have an enhanced ability to acquire novel surface glycoproteins, thus causing public health concerns because of their potential pandemic threat (26).
Although SIVs regularly cause illness in pigs and do not commonly infect humans, sporadic cases of human infection have been reported in many countries (10,18,20,23,31). Between 1990 and 2010, a total of 26 cases of human infections with SIVs were detected in the United States, most of which were isolated single cases of infection, with a history of direct or indirect exposure to pigs prior to onset of illness. In two cases, limited human-to-human transmission could not be excluded (3,4,23). Because sporadic human infections of SIVs were detected through passive surveillance for human influenza, for which only a small subset of influenza A-positive samples are comprehensively analyzed, it was unclear how many human cases of SIV were occurring undetected. Thus, real-time reverse transcriptase PCR (rRT-PCR) procedures were designed and optimized at the Influenza Division of the Centers for Disease Control and Prevention (CDC) in 2008 to allow for rapid, sensitive, and specific detection of these cases (25). In April 2009, the first human cases of 2009 A (H1N1) pandemic (pdm) influenza were detected using the CDC rRT-PCR procedures originally designed to detect human cases of tr-SIV infection (12,25). This novel influenza virus, first found in the United States and Mexico (2,7,9,19), had never been previously isolated. Complete genome analysis determined that the 2009 A (H1N1) pdm influenza virus inherited its genes from an N. Am tr-SIV H1N1 virus and a Eurasian "avian-like" SIV (2,7,9,19) (Fig. 1).
This paper describes the development and analytical performance of the CDC rRT-PCR Swine Flu Panel, based on the initial CDC rRT-PCR procedures for detection of N. Am tr-SIV. In response to the influenza pandemic of 2009, the CDC rRT-PCR Swine Flu Panel and required reagents and supplies were distributed globally and implemented for rapid testing of specimens for the 2009 A (H1N1) pdm influenza virus.

Influenza viruses and clinical specimens.
Influenza viruses tested in this study were grown to high titers in either Madin-Darby canine kidney (MDCK) cells or embryonated chicken eggs (ECE) (27). Infectious virus in culture supernatants or allantoic fluids was measured by determination of the number of 50% tissue culture infectious doses per milliliter (TCID 50 /ml) or the number of 50% egg infectious doses per milliliter (EID 50 /ml), respectively (21).
Clinical specimens (nasal washes, nasal swabs, nasopharyngeal swabs, throat swabs, and lower respiratory tract specimens) included in this study were received from U.S. state and local public health laboratories as well as foreign laboratories between April 2009 and June 2010.
RNA extraction. Viral RNA was extracted from 100 l of virus isolates or clinical specimens and eluted in 100-l volumes by the use of a MagNA Pure Compact RNA isolation kit (RNA_Tissue_V3_1 protocol) and a MagNA Pure Compact instrument (Roche Applied Science), according to the manufacturer's instructions.
CDC rRT-PCR Swine Flu Panel primers and probes. The CDC rRT-PCR Swine Flu Panel includes four oligonucleotide primer and probe sets (universal influenza A [InfA], swine influenza A [swInfA], swine H1 [swH1], and RNase P [RP]) designed for the detection and characterization of 2009 A (H1N1) influenza viruses in human specimens. The assay was designed such that all three assays (InfA, swInfA, and swH1) must be positive to indicate detection of N. Am SIV H1N1 or 2009 A (H1N1) pdm virus. A result where only one or two of the three assays are positive would be inconclusive and would require further characterization. The InfA assay was designed for universal detection of the matrix (M) gene of all influenza A viruses. The swInfA assay was designed for universal detection of the NP gene of all N. Am SIV subtypes and for differentiation of the SIV subtypes from those of human seasonal influenza A (H1N1) and A (H3N2) viruses. The swH1 assay was designed to specifically detect the HA gene of classical SIV or N. Am tr-SIV bearing classical H1 subtype as well as 2009 A (H1N1) pdm influenza virus and differentiate them from those of human seasonal influenza A (H1N1). The RP assay detects the human RNase P gene and is used with human clinical specimens to measure the quality of the specimens as well as to indicate that nucleic acid was extracted adequately from the clinical specimen (8).
Oligonucleotide primers and probes were designed based on available nucleotide sequence data from the GenBank database of the National Center for Biotechnology Information (NCBI), NIH, the influenza sequence database of Los Alamos National Laboratories (LANL), and the Global Initiative on Sharing Avian Influenza Data (GISAID). Nucleotide BLAST search (NCBI) analysis was used to verify primer and probe sequence specificity and avoid potential nonspecific reactivity. The primers and probes were designed to have annealing temperatures of approximately 60°C and 70°C, respectively, using PrimerExpress 3.0 software (Applied Biosystems, Foster City, CA). Dual-labeled hydrolysis probes were labeled at the 5Ј end with the reporter dye 6-carboxyfluorescein (FAM) and quenched with Blackhole Quencher 1 (BHQ1) either at the 3Ј end or internally at a modified "T" residue with Spacer3 (3-Sp3) at the 3Ј end to prevent probe extension by Taq polymerase (Biosearch Technologies, Inc.). Primers and dually labeled TaqMan hydrolysis probes were synthesized by Bio-search Technologies, Inc. (Novato, CA) (see Table S1 in the supplemental material) (32).
rRT-PCR conditions. Reaction conditions for rRT-PCR were based on the Food and Drug Administration (FDA)-cleared CDC Human Influenza Virus Real-time RT-PCR Detection and Characterization Panel (CDC rRT-PCR Flu Panel) (1). The CDC rRT-PCR Flu Panel contains the same primer and probe set for detection of the influenza A virus matrix gene as the CDC rRT-PCR Swine Flu Panel. In order to determine the optimal annealing and extension temperatures, thermal gradient analysis was performed in triplicate using the Bio-Rad CFX96 system and extracted RNA from 2009 A (H1N1) pdm influenza virus strain A/California/07/2009 (Table 1; see also Fig. S1 in the supplemental material). Thermocycling rRT-PCR conditions were as follows: 50°C for 30 min, Taq activation for 2 min at 95°C, and 45 cycles of 95°C for 15 s and an annealingextension step with temperatures ranging from 50 to 65°C for 30 s. All three assays showed comparable levels of performance, with annealing temperatures ranging from 50°C to 60°C. The final reaction annealing temperature was set to 55°C, which is 5°C below the maximum optimal annealing temperature (60°C) used to accommodate potential nucleotide mismatches in the primer-probe regions due to virus evolution.
PCR parameters of the CDC rRT-PCR Swine Flu Panel were optimized using Invitrogen SuperScript III Platinum One-Step quantitative RT-PCR kits (Life Technologies) and Stratagene Mx3005P, Bio-Rad CFX96, and Applied Biosystems (AB) 7500 Fast Dx real-time PCR systems. All rRT-PCRs were performed at a total reaction volume of 25 l. Final primer and probe reaction concentrations were 0.8 M and 0.2 M, respectively. All analytical performance data and clinical specimen data were collected using an AB 7500 Fast Dx Real-Time PCR instrument.

CDC rRT-PCR Swine Flu Panel reaction efficiencies.
Reaction efficiencies of the primer and probe sets of InfA, swInfA, and swH1 assays were determined by testing a 5-fold serial   Table 2). The LoD of the CDC rRT-PCR Swine Flu Panel was determined according to the lowest concentration at which all three assays (InfA, swInfA, and swH1) gave positive results. The LoD was determined to be a virus concentration of 10 1.0ϳ1.6 ID 50 /ml for the three viruses tested. This correlates to 10 Ϫ1.3ϳϪ0.7 ID 50 per reaction (5.0 l/reaction). Similarly, quantified synthetic RNA of 2009 A (H1N1) pdm influenza virus (Armored RNA Quant; AsuraGen, Inc.) was tested in duplicate (Table 3) to determine the minimum RNA copy number detectable by the assay. The LoD of the InfA, swInfA, and swH1 assays was found to be 5 copies of RNA per reaction. The cutoff C T value for the CDC Swine Flu Panel was determined to be Ͻ38 for domestic human diagnostic testing purposes under conditions of FDA Emergency Use Authorization in the United States. This cutoff value is based upon data from LoD analysis that are consistent with the cutoff value of Ͻ38 previously established for the CDC rRT-PCR Flu Panel (1).
Analytical specificity and inclusivity. Assay specificity was demonstrated by testing 10 2009 A (H1N1) pdm influenza viruses, propagated in either MDCK cell culture or ECE. Viruses were diluted to approximately 10-fold above the limit of detection of the assay (approximately 10 2 to 10 3 ID 50 /ml), and the extracted RNA was tested in triplicate. As expected, all 10 viruses tested positive in the InfA, swInfA, and swH1 assays at low virus concentrations (Table 4).
To assess the performance of the CDC rRT-PCR Swine Flu Panel with various SIVs isolated from pigs and humans, extracted RNAs from cultured SIVs were tested (Table 5). Two classical SIV H1N1 subtype viruses and five N. Am tr-SIV H1N1 subtype viruses gave positive results in the InfA, swInfA, and swH1 assays. One N. Am tr-H1N2 subtype virus (whose HA originated from a human influenza virus) and one N. Am tr-H3N2 subtype virus were positive in the InfA and swInfA assays and negative in the swH1 assay. One Eurasian-lineage SIV H1N1 subtype virus was InfA positive and swInfA and swH1 negative. All test results were 100% concordant with the expected results.

Monitoring the clinical performance of the CDC rRT-PCR
A number of human clinical specimens received due to inconclusive results that were positive for InfA and swInfA but negative for swH1 were subsequently confirmed as positive for 2009 A (H1N1) pdm influenza virus by bidirectional genetic sequence analysis. Comparison of the nucleotide sequences of the HA gene determined directly from clinical specimens confirmed three nucleotide mismatches within the probe region of the swH1 assay (Fig. 3). However, a consistent pattern of conserved nucleotide mutations whose circulation persisted could not be identified. In fact, among the 7,122 HA genes of 2009 A (H1N1) pdm influenza viruses available from the

DISCUSSION
The CDC rRT-PCR Swine Flu Panel is a highly sensitive and specific assay for detection and characterization of N. Am tr-SIV containing H1 HA subtype as well as 2009 A (H1N1) pdm influenza viruses. The assay was shown to be highly sensitive and detected as low as 10 1.0ϳ1.6 ID 50 /ml of 2009 A (H1N1) pdm or classical tr-SIV H1N1 viruses, which corresponds to approximately 10 Ϫ1.3ϳϪ0.7 ID 50 per reaction. Similarly, the assay was shown to detect as few as 5 copies/reaction of 2009 A (H1N1) pdm influenza virus RNA. Also, the CDC rRT-PCR Swine Flu Panel was demonstrated to be highly specific for characterization of N. Am SIV H1 subtype as well as 2009 A (H1N1) pdm influenza viruses. As expected, all tested classical SIV H1N1 and N. Am tr-SIV H1N1 viruses were positive with all three assays, while tested N. Am tr-SIV H1N2 and N. Am tr-SIV H3N2 viruses were InfA and swInfA positive but swH1 negative, since both of the latter virus subtypes possess HA antigens inherited from human seasonal viruses (13,14,30).
The CDC rRT-PCR Swine Flu Panel did not show any cross-reactivity among all three assays with common non-influenza A virus respiratory pathogens, influenza B cultured viruses, or influenza B virus-positive clinical specimens. Crossreactivity was not observed in the swInfA and swH1 assays used in testing human seasonal influenza A (H1N1 and H3N2) viruses. No cross-reactivity was observed in the swH1 assay when testing for highly pathogenic avian influenza (HPAI) H5N1 viruses and other animal influenza viruses of different subtypes. However, the swInfA assay demonstrated cross-reactivity with HPAI H5N1 viruses and other non-SIV animal influenza A viruses tested in this study due to similarities of primer and probe sequences to those of some of these viruses. Since none of the non-SIV animal influenza virus tested were positive for all three assays on the CDC rRT-PCR Swine Flu Panel, these results would be considered inconclusive and require further genetic analysis. Conclusive positive results were not observed with a broad range of influenza viruses of nonswine origin as well as with other respiratory pathogens, thus demonstrating 100% analytical specificity of the CDC rRT-PCR Swine Flu Panel.
Soon after the first confirmation ( It is well established that influenza viruses evolve rapidly and are capable of adapting and generating diversity through genetic mutation and genetic reassortment, particularly following the emergence of a novel pandemic strain in humans (15,16,28). As the mutation rate and reassortment capabilities of the 2009 A (H1N1) pdm influenza viruses were initially unknown, it was imperative to monitor the performance of the CDC rRT-PCR Swine Flu Panel during the pandemic to determine the effect of any genetic changes due to mutation or reassortment that might affect assay performance. Therefore, U.S. public health laboratories were advised to submit any clinical specimens with inconclusive results to the CDC Influenza Division for further testing and characterization. Among those submitted, six human specimens contained 2009 H1N1 LAIV, giving positive test results by the InfA and swH1 assays but negative results by the swInfA assay. This reactivity was expected, since the 2009 H1N1 LAIV is a reassortant virus that contains six internal gene segments derived from the coldadapted master donor virus, A/Ann Arbor/6/1960 (H2N2), and the HA and NA gene segments from the recommended vaccine virus A/California/07/2009 (H1N1) pdm (5).
Genetic analysis of clinical specimens received during the pandemic by the CDC due to inconclusive results obtained at local laboratories revealed some viruses with various patterns of three mutations in the binding region of the swH1 probe that caused decreased reactivity of the swH1 assay. However, viruses with similar mutations accounted for only 0.11% of sequences submitted to the GISAID database as of September 2010, thus indicating that the vast majority of 2009 A (H1N1) pdm influenza viruses were detectable by all three assays in the CDC rRT-PCR Swine Flu Panel.
The CDC rRT-PCR Swine Flu Panel served as an effective emergency tool throughout the pandemic for rapid detection of 2009 A (H1N1) pdm virus in many countries and was essential for emergency response implementation. Efforts to further update and optimize procedures and reagents for detection of 2009 A (H1N1) pdm influenza viruses in human clinical samples are still necessary in order to address the potential risk of emerging variants with similar reactivity characteristics, as well as to eliminate cross-reactivity of the swInfA assay with avian and animal viruses.